我有一个包含以下表格的数据库:
trips
stores
trips_stores (this table records which stores are visited on each trip)
每次访问都有一个visitStatus:
NULL - Not visited
0 - Visit pending
1 - Visited
2 - Skipped
我只想为每个商店返回一条记录(即使每个商店可能有多次访问)。但是,我只想返回最小的visitStatus数字:NULL,0、1或2。
例如,我当前的查询返回以下内容:
+---------+--------------------------------------+------------+------------+
| storeID | storeCity | visitStatus| tripDate |
+---------+--------------------------------------+------------+------------+
| 634 | BLACKBURN | 1 | 05/17/2014 |
| 634 | BLACKBURN | 0 | 07/30/2014 |
| 634 | BLACKBURN | 2 | 07/30/2014 |
| 636 | WARRINGTON | NULL | 07/30/2014 |
+---------+--------------------------------------+------------+------------+
但我需要它返回:
+---------+--------------------------------------+------------+------------+
| storeID | storeCity | visitStatus| tripDate |
+---------+--------------------------------------+------------+------------+
| 634 | BLACKBURN | 0 | 07/30/2014 |
| 636 | WARRINGTON | NULL | 07/30/2014 |
+---------+--------------------------------------+------------+------------+
我尝试了GROUP BY和ORDER BY的各种组合,但我不知道。任何帮助表示赞赏
我目前的查询
SELECT
stores.id AS storeID,
stores.city AS storeCity,
trips_stores.status AS visitStatus,
trips.trip_date AS tripDate
FROM stores LEFT JOIN trips_stores on stores.id = trips_stores.store_id
LEFT JOIN trips ON trips_stores.trip_id = trips.id
WHERE stores.map_display = 1
ORDER BY storeID ASC
最佳答案
您需要group by
进行此查询或更复杂的查询。使用group by
,可以使用substring_index()
/ group_concat()
技巧:
SELECT s.id AS storeID, s.city AS storeCity,
(case when sum(ts.status is null) > 0 then NULL
else min(ts.status)
end) AS visitStatus,
substring_index(group_concat(t.trip_date order by coalesce(ts.status, -1)), ',', 1
) AS tripDate
FROM stores s LEFT JOIN
trips_stores ts
on s.id = ts.store_id LEFT JOIN
trips t
ON ts.trip_id = t.id
WHERE s.map_display = 1
GROUP BY storeid, storecity
ORDER BY storeID ASC;
正确处理
NULL
值需要一些逻辑。因为NULL
是最小值-即使存在其他值,也需要在min()
中明确检查。